MATLAB: What is the problem with this code for pitch control of dfig

dfig

L=0.01:0.1:15; c1=0.5176; c2=116; c3=0.4; c4=5; c5=21; c6=0.0068; pitch=0:5:25; for i=1:6 for p=1:length(L); A(p)=1/(L(p)+0.08*pitch(i))-0.035/(pitch(i)^3+1); C(p)=c1*(c2*A(p)-c3*pitch(i)-c4)*exp(-c5*A(p))+c6*L(p); end plot(A(p),C(p)); hold on; end axis ([0 15 -0.1 0.5]); xlabel('\lambda'),ylabel('Cp');

I want the results attached below but i am getting empty screen also attached below. help me

Best Answer

clc; clear all ;
%given matrix
L=0.01:0.1:15; 
c1=0.5176; 
c2=116; 
c3=0.4; 
c4=5; 
c5=21; 
c6=0.0068; 
pitch=0:5:25; 
A = zeros(1,length(L)) ;
C = zeros(1,length(L)) ;
for i=1:6 
    for p=1:length(L); 
        A(p)=1/(L(p)+0.08*pitch(i))-0.035/(pitch(i)^3+1); 
        C(p)=c1*(c2*A(p)-c3*pitch(i)-c4)*exp(-c5*A(p))+c6*L(p); 
    end
    plot(A,C); 
    hold on; 
end
axis ([0 15 -0.1 0.5]); 
xlabel('\lambda'),
ylabel('Cp');

Related Solutions

MATLAB: Runge-kutta 4th order Method Help

All of your F's need to be functions that operate on the current full states. E.g., take this one

F_C1 = (xc*yc*((C1.*R1)./(R1+r0)))-((p.*C1./(p.*C1+(1-p).*C2)).*xp.*yp.*((P.*C1)/(c0+C1)))-xc.*C1;

It is obvious that the C1 derivative depends on the current states of C1, C2, R1, and P. So you need a function that takes these as inputs to compute the derivative. E.g., something like this

F_C1 = @(t,C1,C2,R1,P)(xc*yc*((C1.*R1)./(R1+r0)))-((p.*C1./(p.*C1+(1-p).*C2)).*xp.*yp.*((P.*C1)/(c0+C1)))-xc.*C1;

So modify your code to turn all of those F's into function handles as shown above.

Then you need to realize that all of your variables are interdependent, so using predictor methods will need to predict all of the states simultanously, not one at a time as you are coding. E.g., take these lines in your current code for C1:

    kC1_1 = F_C1(t(i),C1(i));
    kC1_2 = F_C1(t(i)+0.5*h,C1(i)+0.5*h*kC1_1);
    kC1_3 = F_C1((t(i)+0.5*h),(C1(i)+0.5*h*kC1_2));
    kC1_4 = F_C1((t(i)+h),(C1(i)+kC1_3*h));

The first line would change to this using our function handle from above:

    kC1_1 = F_C1( t(i), C1(i), C2(i), R1(i), P(i) );

But what about the next line? It would have to change to something like this:

    kC1_2 = F_C1( t(i)+0.5*h, C1(i)+0.5*h*kC1_1, C2(i)+0.5*h*kC2_1, R1(i)+0.5*h*kR1_1, P(i)+0.5*h*kP_1 );

where the C1, C2, R1, and P predictions are using their respective kC1_1 and kC2_1 and kR1_1 and kP_1 values. But some of these values haven't been computed yet for the current states, such as kP_1. Similar problems will exist in your other derivative calculations. Soooo ... you need to rearrange your code to compute all of the k_1 values first, then compute all of the k_2 values next, etc. E.g.,

kR1_1 = etc.
kR2_1 = etc.
kC1_1 = etc.
kC2_1 = etc.
kP_1 = etc.
kR1_2 = etc.
kR2_2 = etc.
kC1_2 = etc.
kC2_2 = etc.
kP_2 = etc.
   :
 etc.

Give a shot at making these changes and then come back to us with any further problems you might have.

Finally, at some point you will probably get tired of making lots of copy & paste of the RK4 code for each variable. One thing to look into is making your state a 5-element vector instead of using five different variables. That way you only need to write one set of RK4 code that will apply to this vector. This also makes it so that you can use the built-in functions such as ode45( ) to compare with.

MATLAB: Matrix values from loop to raw vector

I've fixed your code:

clear;
q=0.1;
L=2.5;%m

 p=0.52;
c1=0;
c2=(1:10); 
C=zeros(4,40);
for n=(1:10)
    C(:, 4*(n-1)+1:4*n) = [c2(n)+c2(n), -c2(n), -c2(n), (p*c2(n)-(1-p)*c2(n))*L; -c2(n), c1+c2(n), 0, -p*c2(n)*L; -c2(n), 0, c1+c2(n), (1-p)*c2(n)*L;(p*c2(n)-(1-p)*c2(n))*L, -p*c2(n)*L, (1-p)*c2(n)*L, p^2*L^2*c2(n)+(1-p)^2*L^2*c2(n)];
end

Note 1: Use comma instead of space for separator.

Note 2: I indexed matrix C in a different way.

If you want to "create a vector of TEN 4x4 matrices", then use cell object. Later you can still convert it to matrix form using function cell2mat() if you want to.

clear;
q=0.1;
L=2.5;%m
 p=0.52;
c1=0;
c2=(1:10); 
C=cell(1,10);
for n=(1:10)
    C{1,n}=[c2(n)+c2(n), -c2(n), -c2(n), (p*c2(n)-(1-p)*c2(n))*L; -c2(n), c1+c2(n), 0, -p*c2(n)*L; -c2(n), 0, c1+c2(n), (1-p)*c2(n)*L;(p*c2(n)-(1-p)*c2(n))*L, -p*c2(n)*L, (1-p)*c2(n)*L, p^2*L^2*c2(n)+(1-p)^2*L^2*c2(n)];
end
M = cell2mat(C);

Best Answer

Related Solutions

MATLAB: Runge-kutta 4th order Method Help

MATLAB: Matrix values from loop to raw vector

Related Question